EP1410028B1

EP1410028B1 - Immobilisation of proteins and electroactive compounds on a sensor surface by means of a hydrophobin coating

Info

Publication number: EP1410028B1
Application number: EP02743968A
Authority: EP
Inventors: Ewa Maria Rogalska; Renata Bilewicz; Denis Etienne Marie André TAGU; Alain Georges Ghislain Walcarius; Johannes Wemer; Karin Scholtmeijer; Rick Rink; Harm Jan Hektor
Original assignee: Applied Nanosystems BV
Current assignee: Applied Nanosystems BV
Priority date: 2001-07-23
Filing date: 2002-06-21
Publication date: 2008-12-10
Anticipated expiration: 2022-06-21
Also published as: EP1279742A1; WO2003010331A3; DE60230276D1; EP1410028A2; US20040224137A1; US7393448B2; AU2002345432A1; WO2003010331A2; ATE417272T1

Abstract

The present invention relates to a method of binding a compound to a sensor surface, said method comprising the step of adsorbing hydrophobin to said sensor surface. According to the invention, the compound is chosen from the group consisting of i) an electroactive compound, an ii) a compound capable of being converted into an electroactive compound, said method comprising the steps of e) coating the electrode with a hydrophobin, and f) contacting the compound with the hydrophobin to form a hydrophobin coating containing said compound in a non-covalently bound form.

Description

The present invention relates to a method of binding a compound to at least a part of a sensor surface (which part preferably comprises an electrode), said method comprising the step of adsorbing a hydrophobin- to said sensor surface. Classically, hydrophobins are a class of small secreted cysteine-rich proteins of fungi or proteins of bacteria that assemble into amphipathic films when confronted with hydrophilic-hydrophobic interfaces. Some hydrophobins form unstable, others extremely stable, amphipathic films. By assembling at a cell wall-air interface some have been shown to provide for a hydrophobic surface, which has the ultrastructural appearance of rodlets as on aerial hyphae and spores. Some hydrophobins have been shown to assemble into amphipathic films at interfaces between water and oils, or hydrophobic solids, and may be involved in adherence phenomena. It appears that hydrophobins are among the most abundantly produced proteins of fungi, and individual species may contain several genes producing divergent hydrophobins, possibly tailored for specific purposes. Hydrophobins have now been implicated in various developmental processes, such as formation of aerial hyphae, fruit bodies and conidia, and may play essential roles in fungal ecology, including spore dissemination, pathogenesis and symbiosis. Hydrophobins fulfill a broad spectrum of functions in fungal growth and development. For instance, they are involved in formation of hydrophobic aerial structures (e.g. aerial hyphae and fruiting bodies) and mediate attachment of hyphae to hydrophobic surfaces resulting in morphogenetic signals. The mechanisms underlying these functions is based on the property of hydrophobins to self-assemble at hydrophilic-hydrophobic interfaces into amphipathic films. Hydrophobins secreted by submerged hyphae will diffuse in the aqueous environment and may self assemble at the interface of the medium and the air. This is accompanied by a huge drop in water surface tension, enabling hyphae to breach the interface and to grow into the air. On the other hand, hydrophobins secreted by hyphae that contact a hydrophobic environment will self-assemble at the hyphal surface. The hydrophilic side of the amphipathic film interacts with the hydrophilic polysaccharides of the cell wall, while the hydrophobic side becomes exposed to the hydrophobic environment. Aerial hyphae and spores thus become hydrophobic, while hyphae that grow over a hydrophobic substrate firmly attach to it. Hydrophobins are thus active in the environment of the fungus and at the hyphal surface. Moreover, they also function within the matrix of the cell wall where they somehow influence cell wall composition. In this case monomeric rather than self-assembled hydrophobin seems to be involved.
The best characterised class I hydrophobin is SC3 of Schizophyllum commune but, as far as is known by testing, other members of this class have similar properties. Upon contact with hydrophilic-hydrophobic interfaces, SC3 monomers self-assemble into a 10 nm thick amphipathic film. The hydrophilic and hydrophobic sides of the SC3 membrane have water contact angles of 36° and 110°, making these sides moderately hydrophilic (comparable to carbohydrate) and highly hydrophobic (comparable to Teflon), respectively. Interfacial self-assembly of SC3 involves several conformational changes. β-Sheet rich monomers initially adopt a conformation with increases α-helix (α-helix state). SC3 is arrested in this intermediate state at the water-Teflon interface but at the water-air interface the protein proceeds to a form with increased β-Sheet. Initially, this so-called β-Sheet state has no clear ultrastructure (β-Sheet I state) but after a few hours a mosaic of bundles of 10 nm wide rodlets is observed (β-Sheet II state). This ultrastructural change is not accompanied by a detectable change in secondary structure. The transition from the α-helix state to β-Sheet state can also occur at a water-solid interface but has to be induced by increasing the temperature and by adding detergent. Upon self-assembly the properties of hydrophobines change. Hydrophobines in the β-Sheet state is highly surface active, while monomers have no detectable surface activity. Moreover, lectin activity is increased. In addition, the α-helix state form appears to be less stable than the β-sheet state. Although both forms strongly adhere to hydrophobic surfaces, the α-helix form can be dissociated and converted to the monomeric formation by treatment with cold diluted detergents. In contracts, the conformation of the β-Sheet form and its interaction with the hydrophobic solid is not affected by this treatment. Hydropobins, whether or not chemically or genetically modified, can be used to change the biophysical properties of a surface. In this way, the binding of molecules or cells to surfaces could be controlled. For instance, the binding of pathogenic bacteria to catheter surfaces could be reduced while the binding of human fibroblasts to implant surfaces could be encouraged. Apart from changing the biophysical properties, hydrophobins could be used to attach molecules to surfaces that they do not normally have a high affinity with. Attachment could be achieved by chemical cross-linking after the hydrophobin has been assembled on the surface. For example, proteins could be attached to the mannose residues at the hydrophilic side of assembled SC3 via a Schiff-base reaction. Alternatively, in the case of proteins or peptides, fusion proteins can be made and assembled on the surface of interest.
The term "hydrophobin " as used herein comprises not only hydrophobins as isolated from nature and substantially free of other fungal substances such as carbohydrate polymers like schyzophylan, but also includes substances that can be obtained by chemically modifying classically known hydrophobins or by genetically codifying hydrophobin genes to obtain genetically modified proteins not at present available from nature, still having the desired amphipathic characteristics. Classically known hydrophobins (see for example WO 96/41882 which also provides guidance to obtain genetically modified hydrophobin-like substances) commonly are proteins with a length of up to 125 amino acids, with a conserved sequence X_n-C-X_5-9-C-C-X_11-99-C-X_8-23-C-X_6-9-C-C-X_6-18-C-X_m wherein X, of course, represents any amino acid, and n and m, of course, independently represent an integer as disclosed by Wessels et al. (ref. 8). Most classical hydrophobins contain the eight conserved cysteine residues that form four disulphide bridges. However, when the disulphide bridges of a hydrophobin are reduced by chemical modification and the sulfhydryl groups blocked with for example iodoacetamide the protein assembles in water in the absence of a hydrophilic-hydrophobic interface. The structure is indistinguishable prom that of native hydrophobin assembled at the water-air interface. Apparently, the disulphide bridges of hydrophobins keep monomers soluble in water e.g. within the cell in which they are produced or in the medium, allowing self-assembly at a hydrophilic-hydrophobic interface but are not necessary to provide for its amphiphatic character per se.
Class I and class II hydrophobins are known, each at about 100 amino acids in length, having characteristic hydropathy patterns. Most, but not all, contain eight conserved cysteine residues that form intramolecular disulphide bridges. Hydrophobins may be glycosylated, but the characteristic amphipathic properties of these proteins can be solely attributed to their amino acid sequences Although the amino acid sequences of class II hydrophobins are relatively well conserved, those of the class I hydrophobins show a low homology. It would be hard, if not impossible, to design universal primers to pick up class I hydrophobin genes by for example polymerase chain reaction.
Indeed, all hydrophobins that have been physically isolated self-assemble at hydrophilic-hydrophobic interfaces into amphipathic membranes. One side of the hydrophobin membrane is moderately to highly hydrophilic (water contact angles below 90°, for example ranging between 22° and 63°), while the other side exposes a surface with water contact angles essentially above 90°, for example ranging between 93° and 140°, for example as hydrophobic as Teflon (polytetrafluorethylene) or paraffin (water contact angle at about 110°). The contact angle can be determined by a Drop Shape Angle Analysis System, for example the DSA10MK2 supplied by Krüss. The membranes formed by class I hydrophobins (e.g. those of SC3 and SC4 of S. commune) are highly insoluble [resisting 2% sodium dodecyl sulphate (SDS) at 100°C] but can be dissociated by agents such as formic acid (FA) or trifluoroacetic acid (TFA). In contrast, membranes of the class II hydrophobins cerato-ulmin (CU) of Ophiostoma ulmi and cryparin (CRP) of Cryphonectria parasitica readily dissociate in 60% ethanol and in 2% SDS, while assembled CU is also known to dissociate by applying pressure or by cooling. Self-assembly of hydrophobins is accompanied by conformational changes. Monomeric class I and class II hydrophobins are rich in β-sheet structure. At the water-air interface, class II hydrophobins attain more β-sheet structure (called the β-sheet state), while at the interface between water and hydrophobic solid, a form with increased α-helix is observed (the α-helical state). The α-helical state seems to be an intermediate of self-assembly, whereas the β-sheet state is likely the stable end-form. At the water-air interface, monomers of class I hydrophobins attain the α-helical state within seconds, but the conversion to the β-sheet state is much slower and takes minutes to hours. At the water-solid interface, the protein also readily attains the α-helical state but is thought to be arrested in this intermediate state. The β-sheet end state can then be reached by applying a combination of heat and diluted detergent. Both forms of the assembled hydrophobin have an amphipathic nature and can be dissociated with TFA, which unfolds the protein. After removing the solvent and dissolving in water, class I hydrophobins refold to the same monomeric structure that was observed before purification or TFA treatment. However, self-assembly and disassembly of class II hydrophobins can also be repeated even after dissociation of the membrane by TFA. This shows that both classes of hydrophobins are highly resilient to this type of treatment. The membrane of class I hydrophobins is characterized by a mosaic of bundles of 5-12 nm-wide parallel rodlets. In contrast, rodlets have not been found at surfaces of the assembled class II hydrophobins CFTH1 of Claviceps fusiformis, CRP of C. parasitica, and HFB1 and HFB2 of Trichoderma reesei. Whether the absence of rodlets or the differences in rodlet diameter has any functional significance is not yet known. The rodlets of the class I hydrophobins, SC3 and SC4, of Schizophylum commune are very similar to the fibrils formed by amyloid proteins. They consist of two tracks of 2-3 protofilaments with a diameter of about 2.5 nm each, have a high degree of β-sheet structure, and interact with the fluorescent dyes Thioflavine T (ThT) and Congo Red. Both dyes can be used as probes to discriminate between the alpha-helical state and the beta-sheet state, each having a high propensity for beta-sheet state but no or lower propensity for alpha-helix state or soluble hydrophobin. In addition, SC3 and amyloid proteins self-assemble via intermediates and only above a critical concentration. It was suggested that amyloid fibril formation is common to many, if not all, polypeptide chains. However, because formation of amyloid fibrils is accompanied by loss of function or even disease (e.g. Alzheimer's disease), evolution would have selected against the propensity to form such fibrils. Yet, one or two mutation(s) in a protein suffice to considerably increase the tendency to form amyloid fibrils. To our knowledge, hydrophobins are the first example of functional amyloids, with multiple functions in fungal development. Recently, it was found that the four disulfide bridges of the SC3 hydrophobin are essential to prevent the protein from forming the amyloid structures in the absence of a hydrophilic-hydrophobic interface. When the disulphide bridges were reduced and the sulfhydryl groups blocked with iodoacetamide, the protein spontaneously assembled in water. Its structure was then indistinguishable from that of native SC3 assembled at the water-air interface. Apparently, the disulphide bridges of hydrophobins keep monomers soluble in water (e.g. within the cell or in the medium) and thus prevent precocious self-assembly. This would explain why in nature most hydrophobin have eight conserved cysteine residues.
Hydrophobins belong to the most surface-active molecules. With a maximal lowering of the water surface tension from 72 to 24 mJ m^-2 at 50 µg ml^-1, SC3 is the most surface-active protein known. Other hydrophobins are also highly surfaces active. Their surface-lowering activities are at least similar to those of traditional biosurfactants. In contrast to these surfactants, surface activity is not dependent on a lipid conjugate but is solely caused by the amino acid sequence. Moreover, while the maximal lowering of the surface tension by the traditional surfactants is attained within seconds, it takes minutes to hours in the case of class I hydrophobins. This is explained by the fact that hydrophobins lower the water surface only after self-assembly that is accompanied by conformational changes in the molecule.
Despite the fact that hydrophobins have diverged considerably, their gross properties are similar. This flexibility is also illustrated by the fact that removing 25 out of 31 amino acids preceding the first cysteine residue of the SC3 hydrophobin to generate truncated SC3 by genetic engineering only affected the wettability of the hydrophilic side of the assembled hydrophobin. A most remarkable hydrophobin is the trihydrophobin CFTH1 of C. fusiformis. It contains three class. II hydrophobin-like units, each preceded by a Gly-Asn-rich repeat and still behaves like other class II hydrophobins. Because of the interfacial self-assembly into amphipathic protein films, hydrophobins can change the wettability of surfaces.
As said, one method to measure wettability is by estimating or measuring the contact angle that a water drop makes with the surface. A large contact angle (>90 deg) indicates a hydrophobic, a small contact angle (<90 deg) a hydrophilic surface. Furthermore, in gas/liquid or liquid/liquid systems, such as in vigorously shaken water or in oil-in-water or water-in-oil dispersions, air bubbles or oil droplets in solution of hydrophobin become coated with an amphipathic film that stabilizes them. Solid/liquid interfaces show the same stabilisation. For example, a sheet of hydrophobic plastic such as Teflon (contact angle 110 deg) immersed in hydrophobin becomes coated with a strongly adhering protein film that makes the surface completely wettable (contact angle 48 deg), even after SDS extraction (contact angle 62 deg), and hydrophobin monomers dried down on a hydrophilic surface make the surface hydrophobic.
The classical hydrophobins are i) typically isolated from fungi like (ref. 8) but can now also be made recombinantly; or ii) comprise a polypeptide having at least 40% identity and at least 5% similarity to at least one polypeptide chosen from the group consisting of i) amino acids 29 - 131 of SEQ NO. 1 and ii) amino acids 29 - 133 of SEQ. NO. 2. Such a protein may be derived from a filamentous bacterium, in particular a bacterium capable of forming aerial hyphae such as an and more specifically the filamentous bacterium may be a species. A species from which the protein may be isolated using standard procedures for the isolation of hydrophobins, is a species which has been transformed with a construct that can be isolated from an strain which has been deposited on 14 March, 2000 under accession number CBS 102638 with the Centraalbureau voor Schimmelcultures (Oosterstraat 1, P.O. Box 273, 3740 AG Baarn, the Netherlands). This is disclosed in PCT/NL01/00268 .
In particular, Wessels et al. (Advances in Microbial Physiology, 38, pp. 1-45 (1997)) suggest attaching small ligands to a layer of hydrophobin via covalent binding or coupling of amino groups to aldehyde groups on mannose residues (p. 35). The (only) example given relates to coupling a protein molecule onto a layer of hydrophobin present on a gold surface. However, coating the surface with a hydrophobin may reduce the sensitivity of the sensor, as less surface area is available, or reactions to be detected take place at a greater distance from the sensor surface.
Furthermore, Wösten et al. (Bioch. Bioph. Acta 1469 (2000), pg. 79-86) describes the capacity of hydrophobins to self-assemble at a hydrophilic-hydrophobic interface into an amphipathic film. Discussed are the changes in secondary structure of hydrophobins upon self-assembly. Nothing is mentioned about attaching a compound to the assembled hydrophobin layer.
Pum et al. (Trends Biotechn. 1999, Vol. 17, no.1., pg 8-12) relates to self-assembled monolayers (SAMs) of a distinct group of proteins, namely crystalline bacterial-cell-surface-layer (S-layer) proteins, and the use of SAMs for immobilizing biologically functional molecules on biosensors, e.g. by electrostatic interactions or after carbodiimide activation of the carboxyl groups.
The invention provides a method of providing a sensor surface with a compound, said method comprising the steps of providing at least a part of the surface of the sensor with a hydrophobin coating and contacting the compound with the hydrophobic coating to form

a non-covalent bond between said hydrophobin coating and said compound In one embodiment, as commented upon by Bilewicz et al in J. Phys. Chem. B 2001, 105, 9772-9777 of August 2001 as particularly useful for electrodes, a method according to the present invention is provided in that the compound is chosen from the group consisting of i) an electroactive compounds, an ii) a compound capable of being converted into an electroactive compound, said method comprising the steps of coating the electrode with a hydrophobin
and contacting the compound with the hydrophobin to form a coating containing said compound in a non-covalently bound form.

A particular type of sensor is the electrochemical sensor, comprising an electrode as the sensor surface. Coating an electrode with a hydrophobin would in general result in reduced access of electroactive compounds to the surface of said electrode. Surprisingly, we have found that it is possible to non-covalently incorporate a compound in said hydrophobin coating. The compound remains in the hydrophobin coating for a substantival time, as evidenced by experiment. Such a compound becomes incorporated when it has a lower molecular weight than the hydrophobin, preferably less than 2000, and more preferably less than 1000 dalton. The electroactive nature of the compound improves the sensitivity of the electrodes in comparison with a hydrophobin coating not containing said compound.
Another application of the invention relates to a method of providing a sensor type possessing favorable features for use as a biosensor, which is a device that incorporates a biological recognition element in close proximity or integrated with the signal transducer, to give an essentially reagentless sensing system specific to a target compound. In the method provided in the invention, the biological recognition elements are non-covalently attached to a biosensor surface in a non-covalent manner via a hydrophobin coating. This mode of biding leaves the secondary and tertiary structure of such biological compounds virtually intact and thus allows improved biological recognition compared to covalent binding methods which often alter such structural determinants.
In another embodiment, a method according to the present invention is provided wherein a fist (non-covalently attached) compound is a proteinaceous substance with a higher molecular weight than a classically known hydrophobin (i.e. > 15 kD) but, not-withstanding, becomes non-covalently attached to said coating without essentially losing its reactivity. Surprisingly, substantial non-covalent attachment of such a larger compound to a sensor surface, whereby preferably reactivity of said compound Such as enzymatic activity, or its propensity to bind to a ligand or antigen, is maintrained, can be achieved by immersing said sensor surface in a solution comprising said compound. Where covalent linking is classically thought to be essentially related to the presence of glycosylated hydrophobin, it is now provided that said surfaces can also be provided with larger molecules when said coating is essentially devoid of mannose residues that would allow for such covalent linking. This facilitates the use of other hydrophobins for coating sensor surfaces to which larger molecules may be attached. Where classically one used SC3, which in its natural states is provided with an N-terminal side comprising glycosylated residues, it is also possible to use non-glycosylated substances, such as trSC3 (truncated-SC3) from which said glycosylated N-terminal is absent, but also other hydrophobins characterised by an amphipathic protein character, for example those not having all the classically conserved cysteine-residues in place and essentially capable of providing a hydrophobic surface with a hydrophilic face, or vice versa.
It is in particular herein provided to provide sensor surfaces with molecules of proteinaceous nature, such molecule may comprise a conventional (poly- or monoclonal) or synthetic antibody or other binding molecule (optionally additionally provided with an enzyme, such as a peroxidase), an enzyme, such as glucose oxidase or cholesterol oxidase as provided herein in the detailed description, or alkaline phosphatase, luciferase, esterase, lipase, or trypsin, or combinations of these or other enzymes, a receptor for measuring ligand interaction, a light receptor or light harvesting complex, combinations thereof, and so on. Where covalent binding often results in loss of reactivity, these compounds bound with a method according to the invention in particular substantially maintain their reactivity towards ligands, antigens, substrates, etc, probably exactly because the bond with the coating is of a non-covalent nature.
Herein, the term "identity" used in association with a proteinaceous substance such as a (poly)peptide is defined, in accordance with the state of the art, as having exactly matched amino acid residues. Here, sequences may comprise insertion or deletions. The term "similarity" used in association with proteinaceous substances such as a (poly)peptide denotes conservative substitutions. Conservative substitutions are substitutions in which one amino acid is replaced with another, where the following amino acids are considered similar:

A,S,T;
D,E;
N,Q;
R,K;
I,L,M,V;
F,Y,W.

The term "electro-active" is defined as a compound which can undergo changes in the oxidation state, i.e. undergo a redox reaction. An electro-active compound, such as Q10, azobenzene, Q0 or calixerene, will have a lower molecular weight than the hydrophobin, and more in particular it will have a MW of less than 2000 dalton, and more preferably less than 1000 dalton. Compounds capable of being converted into an electroactive compound are known in the art. They may become electroactive after, for example, irradiation with light. This allows the electroactive compound to be made available at a desired time. Some of the advantages of the use of such a compound are

more accurate measurements (background measurements can be made before the electroactive compound is released)
reduced consumption of electroactive compound. Generally, the compound capable of being converted into an electroactive compound (an example of which is calixarene) will have a molecular weight as specified above.

In particular, the compound is a hydrophobic compound or a compound containing a hydrophobic anchor. Such compounds are among the compounds most stably maintained in the hydrophobin coating. It is thought that a planar hydrophobic compound or anchor may be beneficial. In the present application the term "anchor" is understood to mean a part of the compound, said part having a side and/or moiety lacking hydrophilic groups. It is also thought that the absence or a reduced number of negative and/or positive charges is advantageous. If charge is present, it is preferably from weakly acidic or basic groups, which can release or accept a hydrogenium ion to eliminate the charge.
Advantageously, a second compound is bound covalently to the hydrophobin, the second compound being an electroactive compound.
According to an alternative embodiment, said second compound is bound non-covalently to the hydrophobin, through a third compound being an intermediate compound having affinity for the second compound, said second compound being an electroactive compound.
Both these methods allow for a (more) selective measurement. With the ("first") compound being present, good sensitivity is achieved even though the second compound is at a distance from the electrode surface.
Hence, preferably the second compound is a proteinaceous substance such as redox enzyme or a light receptor. A light receptor may, for example, be a haem-group or a chlorophyl-group.
The invention also relates a sensor comprising to an electrode coated with a hydrophobin the coating incorporating a compound being an electroactive compound or provided with molecules > 15 kD. The sensor is provided with a sensor surface at least a part of which is provided with a hydrophobin coating comprising a non-covalently bound compound. In one embodiment, a sensor according to the invention is provided wherein said compound is smaller than 15000 dalton, preferably smaller than 2000, or even 1000 dalton, as is the case with the electroactive compounds Q10. azobenzene, Q0, calixarene, and others. In another embodiment a sensor according to the invention is provided wherein said compound is larger than 15000 dalton, preferably even larger such as >25 kD, considering the average size of at least the functional parts of enzymes, receptors, antibodies, and the like. It is herein also provided to coat said sensor non-covalently with a large or small proteinaceous substance, such as a peptide or polypeptide. To perform voltometric analysers, it is preferred that such an analytical system comprises an electrode such as a glassy-carbon electrode, a gold electrode or a Thin Mercury Film Electrode, which can advantageously be further provided an electroactive first, second or third compound to improve determining a voltage difference in a solution comprising contacted said solution with a sensor according to the invention and recording a current.
The electrode is manufacture using any embodiment of the method according to the invention. The electrode may be, for example a gold or platinum electrode, and in particular, the electrode is a glassy-carbon electrode (GCE), a glass electrode (GE) or a Thin Mercury Film Electrode (TMFE).
The sensor may comprise further means for conducting a measurement, such as lead wires, integrated or non-integrated devices such as an amplifier, a reference electrode, signal processing means, as is well known in the art.
Furthermore, the invention relates to a method of providing a sensor type possessing features that are favorable for use as a biosensor. Biosensors use biological molecules to detect other biological molecules or chemical substances. Biosensors might for example use a monoclonal antibody to detect an antigen, or a small synthetic DNA molecule to detect DNA. In a biosensor, a biological recognition (sensing) element is in direct spatial contact with a transducer element, to give a reagentless sensing system specific to a target compound (analyte). Transducers are the physical components of the sensor that respond to the products of the biosensing process, which may be optical, electrochemical, thermometric, piezoelectric or magnetic, and outputs the response in a form that can be amplified, stored, or displayed. The biological recognition element may be a biological material or a biomimic (e.g. tissue, microorganisms, organelles, cell receptors, enzymes, antibodies, nucleic acids etc.). Biological sensing elements have a remarkable ability to distinguish between the analyte of interest and similar substances with great accuracy. Biosensing occurs only when the analyte is recognized specifically by the biological element. It is preferred that the biological elements are bound to the sensor surface in a non-covalent manner, as is the case in the method provided in the invention, leaving secondary and tertiary structures of such biological compounds virtually intact and thus allowing optimal biological recognition.
For a given analyte-recognition element reaction, several transduction schemes may be applicable. Amperometric devices detect changes in current as constant potential. Conductimetric devices detect changes in conductivity between two electrodes. Potentiometric devices detect changes in potential at constant current (usually zero). Optical transducers can be subdivided into two modes (extrinsic and intrinsic) according to the optical configuration. In the intrinsic mode, the incident wave is not directed through the bulk sample, but propagates along a wave guide and interacts with the sample at the surface within the evanescent field. Other surface methods of optical detection of biological recognition are based on modulation of the field excited at the interface between different materials due to incident light. For example, the BIAcore system monitors bio-specific-interactions with a surface plasmon resonance detector to detect minor mass changes at the surface such as antibody binding to a surface-immobilized antigen.
A biosensor is distinguished from a bioanalytical system which requires additional processing steps, such as reagent addition. The reagentless form in most cases is achieved by immobilizing the biological recognition element onto the sensor surface. Advantageously, the biosensor surface is coated with a compound allowing non-covalent attachment of a biological recognition molecule. Where covalent binding often results in loss of reactivity, these compounds bound with a method according to the invention in particular substantially maintain their reactivity towards ligands, antigens, substrates, etc, probably exactly because the bond with the coating is of a non-covalent nature. For example, a gold sensor surface (hydrophobic) is coated with a hydrophobin solution by simply incubating for some time. Then the surface is washed with water to remove any unbound hydrophobin. Following treatment of the surface with detergent to obtain a beta-sheet state coating, the coated surface is used directly to bind a biological recognition molecules like antibodies, enzymes, peptides, lipids, nucleic acids or carbohydrates.
In general, the immobilization matrix may function purely as a support, and it is preferred that the immobilization matrix does not interfere with the sensitivity of the biosensor. Whereas the mere coating of a sensor surface with hydrophobin would result in reduced signal transmission, we found that it is possible to incorporate a compound in said hydrophobin coating to increase the sensor performance. For example, as provided herein, such a compounds is an electroactive compound that is incorporated in the hydrophobin coating to improve the sensitivity of the sensor in comparison with a hydrophobin coating not containing said compound. In a preferred embodiment, the invention provides a method of providing a sensor surface with a hydrophobin coating wherein said coating is additionally provided with a reactive compound to improve the sensitivity of the sensor, also referred to as signal transducer, as well as with a non-covalently bound biological recognition compound.
The invention is further explained in the detailed description herein.

Figure legends

Fig. 1a-c show cyclic voltammograms showing the difference between bare (1) and hydrophobin-modified (2) electrodes (GE, GCE. TMFE respectively);
Fig. 2a,b show cyclic voltammograms using ferrocyanate ions as a probe to check the blocking properties of hydrophobin;
Fig. 3a,b are similar to fig. 1 and show the effect of adsorption of ubiquinone Q10 adsorbed to a GC-electrode (GCE);
Fig. 4 shows the effect of diazobenzene on a cyclic voltammogram for a GCE; and
Fig. 5 is similar to fig. 4, except that the effect of ubiquinone Q0 is shown.

Detailed description

Production and purification of the hydrophobin. HYDPt-1 hydrophobin was produced in Escherichia coli as a recombinant polypeptide of 13.7 kDa (ref. 1). Briefly, the hydPt-1 cDNA was cloned in the pQE30 plasmid (Qiagen, Germany) to produce a fusion protein with a His-tag motif. Extraction was performed as described previously (ref. 1). After chromatographic purification on a Ni²⁺ affinity column, the HYDPt-1 polypeptide was concentrated in 10 mM Tris-HCl by ultrafiltration.
Chemicals and Solutions. All chemicals were of analytical grade. Tris(hydroxymethyl)aminomethane (Tris), dimethyloformamide (DMF) and LiOH were from Fluka. Coenzymes Q0 (2,3-dimethoxy-5-methyl-1,4-benzoquinone) and Q10 (ubiquinone 50) were from Sigma. Diazobenzene was from Reachim, Hungary. Methanol, citric acid and K₃[Fe(CN)₆] were from POCh, Poland. All solutions were prepared daily. Distilled water was passed through a Milli-Q water purification system. The surface tension was 72.5 mN m^-1 at 20°C and final resistivity was 18.3 MΩ cm^-1.
Electrochemical experiments. Voltammetry experiments were done at 20°C in three - electrode arrangement, with a calomel reference electrode, platinum foil counter electrode and GE, GCE or TMFE as working electrodes. Eco Chemie AUTO-LAMB PGSTAT 30 system was used as the potentiostat with an IBM PC and Eco Chemie software. In order to prepare the TMFE, a silver wire or silver disk electrode, precleaned in concentrated perchloric acid, was touched to a drop of mercury and cathodically polarized in 0.1 M KOH to obtain a shining and uniform layer of mercury, ca. 1 mm thick. The GE and GCE were polished on Buehler polishing papers and paste, and the GE was then cleaned in concentrated nitric acid. The electrodes were coated with the self-assembled hydrophobin upon contact with the surface of solutions containing 2 µg of hydrophobin per 1 ml of 10 mM Tris-HCl buffer. The time of the hydrophobin self-assembly and the conditions of Q0, Q10 and diazobenzene adsorption on hydrophobin coated electrodes are given below.
Barrier properties of HYDPt-1 films on gold and glassy carbon electrodes. The properties of HYDPt-1 layers adsorbed on hydrophilic and hydrophobic solid surfaces were compared using three different electrode substrates, namely GE, GCE and TMFE. The hydrophilic GE surface was modified with HYDPt-1 by self-assembling the protein at the liquid-air interface, followed by adsorption of the layer to the gold surface. The protein was assembled from 10 mM Tris-HCl buffer, pH 7.0, containing 2 µg/ml hydrophobin, and adsorption to the surface was achieved by lifting the electrode up through the interface, or by a horizontal touching of the hydrophobin-covered water surface with the electrode. Figure 1 shows the cyclic voltammograms recorded using the bare (1) and hydrophobin modified (2) electrodes. Figure 1a allows the comparison of the bare (1) and covered (2) GE. The curves are similar in that no decrease of background current is observed, and no peaks appear in the voltammogram. The presence of the hydrophobin layer on the electrode surface is evidenced by the inhibition of the final increase of anodic current due to gold oxidation. HYDPt-1 is inert in a wide range of potentials and does not lead to a decrease of capacity currents which means that the protein layers formed on the electrode are not as dense and highly blocking, as the layers of, e.g., alkanethiols (ref. 2). The extent of blocking is not changed even after 24 hours of self-assembly.
Wessels and Wösten observed that the SC3 hydrophobin had much higher affinity to hydrophobic than to hydrophilic surfaces (ref. 3, 4). Two types of electrodes, GCE and TMFE, were therefore chosen as model hydrophobic surfaces to check the behavior of HYDPt-1. The results of self-assembly are shown in Figures 1b and 1c respectively. In both cases the background currents become much smaller after modification (2), demonstrating that the coverage of GCE and TMFE with HYDPt-1 is much higher, compared to that of the gold substrate. The protein layers are stable and firmly attached to the electrode substrate, as indicated by the GCE voltammogram which does not change over several weeks. In the case of TMFE, high quality films are formed even when the time of self-assembly is decreased from 24 hours to 20 minutes (fig. 1c). The capacity of the modified TMFE is significantly lowered, and the onset of the mercury oxidation current is shifted towards more positive potentials, revealing strong blocking properties of the hydrophobin layer. An additional peak appears in the TMFE voltammogram at -0.58V. This peak corresponds to the reduction of mercury cysteinate formed on the electrode surface upon oxidation of mercury in the presence of cysteine thiol groups present in the protein.
Stability of HYDPt-1 layers in solutions of different pH. The dependence of stability and blocking properties of HYDPt-1 layers on the pH of the solution was checked by recording multiple cyclic voltammograms using all electrodes in solutions of pH 2.2 (citric acid), pH 4.7 (citric acid / LiOH), pH 7.0, pH 10.2 (Tris), and pH 12.1 (LiOH). The HYDPt-1 layer remained well attached to the electrode surfaces in all solutions studied, and the blocking effect on various substrates followed the behavior observed at pH 7.0.
Probing blocking properties of HYDPt-1 layers using ferrocyanate as the electrochemical probe. The ability of small hydrophilic anions to access the electrode surface through the HYDPt-1 layer was checked using ferrocyanate ions as the electrochemical probe. Cyclic voltammograms were recorded in 0.1 M / HClO₄ solution containing 0.75 mM K₃Fe(CN)₆ (Figure 2). The voltammograms recorded for K₃Fe(CN)₆ using HYDPt-1 coated (2) electrodes are different, compared to those obtained with a bare (1) GE (fig. 2a) or GCE (fig. 2b). The currents are much lower and the voltammetric curves are more sigmoidal in shape. This behavior establishes that the extent of coverage of both electrodes by HYDPt-1 is high, and that the probe has a limited access to the electrode surface. The transition from peaked to sigmoidal shape is expected when the access sites are dispersed (ref. 5, 6), and when spherical diffusion becomes the major process for transporting the molecules to the electrode surface, as distinct from the linear diffusion observed for large bare electrodes. The GCE surface is blocked more efficiently than the gold surface, as shown with the experiments performed in pure supporting electrolyte solution. The latter results confirm a higher affinity of HYDPt-1 towards hydrophobic surfaces.
Hydrophobin as a "molecular glue" for immobilizing molecules on the electrode surface. Wösten and de Vocht suggested that hydrophobin might be used to covalently attach cells to hydrophobic surfaces in medical and sensing devices (ref. 7) via binding to mannose residues. In the present work we checked the ability of HYDPt-1 to bind through sorption different types of electroactive molecules to electrode surfaces. The long hydrocarbon chain ubiquinone (Q10) was used as a model hydrophobic molecule. Firstly, the HYDPt-1 layer was self-assembled on the GCE from the usual solution (2 µg/ 1 ml Tris buffer, pH 7.0). Next, self-assembly of Q10 was carried out from a solution containing 1 mg of Q10 in 1 ml DMF.
Electroreduction of ubiquinone Q 10 immobilized on electrodes modified with HYDPt-1. In neutral aqueous solution ubiquinone undergoes reduction, according to the following scheme:
The voltammetric curve obtained with ubiquinone Q10 adsorbed on the HYDPt-1 modified GE (2 in fig. 3a) and GCE (2 in fig. 3b) is shown. In fig. 3, 1 denotes an electrode covered with hydrophobin only. The shape of the curve and the linear dependence of the peak currents on the scan rate points to surface immobilization of ubiquinone. The GCE substrate covered with HYDPt-1 was found to bind Q10 in a very stable way, giving rise to ubiquinone reduction and oxidation signals which remained unchanged for several weeks.
Electroreduction of diazobenzene immobilized on electrodes modified with HYDPt-1. Diazobenzene is a small molecule with a photo- and electroactive azo group, which does not undergo adsorption on a bare glassy carbon electrode. However, when adsorbed on a HTYDPt-1 modified electrode, diazobenzene remains stably attached to the surface, even after repeated transfers of the electrode into solutions of different pH and not containing the azocompound. Self-assembly of diazobenzene was carried out from a 1 mM methanol solution. Reduction of diazobenzene can be described as shown in the Scheme 2 :
Figure 4 shows the cyclic voltammogram of diazobenzene adsorbed for 20 min on the HYDPt-1 - modified electrode, recorded in 0.1 M Tris / HClO₄ solution of pH 7.0. Curve 1 was recorded after adsorption of the diazobenzene for the same laps of time, but on the bare GCE. Curve 2 represents the electrode covered with hydrophobin, and curve 3 represents the electrode with both hydrophobin and diazobenzene. The well developed reduction and oxidation peaks do not change upon repeated cycling. The peak currents increase linearly with square root of the scan rate, indicating diffusion control rather than surface - immobilized species. Since the working solution does not contain diazobenzene, this dependence can be understood in terms of diffusion of diazobenzene within the HYDPt-1 layer. Such behavior argues that, in the self-assembly process, the small and hydrophobic diazobenzene molecule penetrates into, and is immobilized in the HYDPt-1 layer. The diazobenzene incorporated into the film is now being studied in our laboratories as a molecular switching device, based on the cis-trans isomerization taking place on UV irradiation. Similar scan rate dependencies were observed for the Q0 molecule, which has the same headgroup as Q10 but does not possess an alkyl chain (Figure 5), and therefore can easily penetrate the HYDPt-1 layer. Curve 1 is a bare electrode in the presence of Q0, and curve 2 an electrode covered with hydrophobin and after adsorption of Q0.
Immobilization of glucose oxidase on glassy carbon electrode.Glassy carbon electrode was coated with hydrophobin by placing the electrode in a solution of hydrophobin (100 µg/ml) and incubating for 15 minutes, after which the electrode was thoroughly rinsed with water. Subsequently, the coated electrode was submerged in a glucose oxidase containing solution (SIGMA; 210,000 units/g of solid, final concentration 1.8 mg/ml) for 2 hours, and rinsed with water afterwards. The electrode, after modification and functionalization with the enzyme, was used in a three electrode system, including an Ag/AgCl reference electrode and a Pt counter electrode.Phosphate buffer pH 7 (25 mM) was used as electrolyte. When glucose was added, the immobilized glucose oxidase catalyzed the reaction leading to formation of hydrogen peroxide, which could be detected as a small current, which was proportional to the glucose concentration. The glucose oxidase remained active upon immobilization on the hydrophobin layer.
The modified electrode was stored in a sealed container, not in liquid, at 4 °C, and tested frequently for activity and response to glucose. Over the period tested, 67 days, the electrode maintained its activity and was not influenced by the frequent testing. Immobilization of cholesterol oxidase on glassy carbon electrode. Glassy carbon electrode was coated with hydrophobin by placing the electrode in a solution of hydrophobin (100 µg/ml) and incubating for 15 minutes, after which the electrode was thoroughly rinsed with water. Subsequently, the coated electrode was submerged in a cholesterol oxidase containing solution (0.5U/ml) for 2 hours, and rinsed with water afterwards.
The electrode, modified and functionalized with the enzyme, was placed in a three electrode system, including an Ag/AgCl reference electrode and a Pt counter electrode, using phosphate buffer pH 7 (25 mM) as electrolyte. To overcome the poor solubility of cholesterol in water, the cholesterol was solubilized in isopropanol with Triton in phosphate buffer (Ropers, M-H. et al., 2001, Phys. Chem. Chem. Phys. 3:240-245). Upon addition of cholesterol, the immobilized cholesterol oxidase catalyzed the reaction leading to formation of hydrogen peroxide. The peroxide resulted in a small detectable current over the electrode, which was correlated to the cholesterol concentration added. This example demonstrated the immobilization of cholesterol oxidase, while remaining active.Immobilization of Light Harvesting complex (LHC).
Three different electrode substrates (hydrophilic GE and hydrophobic GCE or TMFE) were coated with hydrophobin (either SC3, TrSC3 or SC4) in a 100 µg/ml hydrophobin solution and were placed in water. This was incubated at 25 °C for different time periods varying from 10 minutes to 16 hours. The electrodes were washed with water to remove any unbound hydrophobin; these coatings were referred to as the α-helix state. One hydrophobin coated and one bare electrode of each type was boiled in 2% SDS-solution for 10 min. The SDS-treated electrodes were extensively washed with water. These coatings were referred to as the β-sheet state. The various types of coated and bare electrodes were loaded with electroactive compounds Q10, azobenzene or Q0 as described by Bilewicz et al in J. Phys. Chem. B 2001, 105, 9772-9777 or with a mediator such as methylene blue. The various types of coated and bare electrodes either or not loaded with the electroactive compounds were incubated in LHC of Cyclotella cryptica, isolated as indicated in Rhiel et al. (Rhiel E. et al., 1997, Botanica Acta 110, 109-117), at various concentrations for 2 h at 25°C. The electrodes were washed with the appropriate buffers. To assay the activity of the immobilized LHC, the electrodes (in the appropriate buffer) were placed in the dark followed by placing them in daylight and measuring the current. The dark-light cycles were repeated several times on the same day, after 1 day, after 1 week and after 1 month to determine the stability of the immobilized LHC.
Immobilization of CYP2D6 and CYP2C19 on an electrode surface. A glassy carbon electrode was coated with hydrophobin by emerging the electrode for 15 min in a solution containing hydrophobin (100 µg/ml). The electrode was thereafter thoroughly rinsed. The cytochromes CYP2D6 and CYP2C19 were separately bound on one hydrophobin coated electrode each, by incubation CYP2D6 or CYP2C19 containing solution for 15 minutes and extensively rinsed. The modified electrodes were placed in a medium containing NADPH and the model substrates dextromethorphan and mephenytoin. During the incubation period the potential was measured, which was induced by the contact with the substrates dextromethropan and mephenytoin. The magnitude of the potential reflects the metabolism by the iso-enzymes. After the incubation for 1 hour, the electrodes were removed and the substrates (dextromethorphan and mephenytoin, respectively) and products (dextrorphan and 4-hydroxymephenytoin, respectively) were quantified. The ratio of both was a reflection of the activity of the cytochromes and was correlated to the measured potential.

REFERENCES

1. Tagu, D., et al. New Phytol. 2001, 149, 127-135;
2. Finklea, H. O. Electrochemistry of Organized Dekker: New York, 1996; Vol 139, pp 109-235;
3. Wösten, H. A. B.et al. EMBO J. 1994, 13, 5848-5854;
4. Wösten, H. A. B. et al. Colloids Surf. B: Biointerfaces 1995, 5,189-194.;
5. Amatore, C. et al. J. Electroanal. Chem. 1983, 147, 39-51;
6. Bilewicz, ;
7. Wösten, H. A. B. et al. Curr. Biol., 1999, 9, 85-88.;
8. Wessels, J. G. H. Adv. Microb. Physiol. 1997, 38, 1-45.

SEQUENCE LISTING

(Source: PCT/NL01/00268 )

<110> Applied Nanosystems B.V.
<120> Protein capable of self-assembly at a hydrophobic-hydrophylic interface
<130> GB44441
<140>
<141>
<160> 4
<170> PatentIn Ver. 2.1
<210> 1
<211> 131
<212> PRT
<213> Artificial Sequence
<220>
<221> SIGNAL
<222> (1)..(28)
<400> 1
<210> 2
<211> 133
<212> PRT
<213> Streptomyces coelicolor
<220>
<221> SIGNAL
<222> (1)..(28)
<400> 2
<210> 3
<211> 396
<212> DNA
<213> Streptomyces coelicolor
<400> 3
<210> 4
<211> 402
<212> DNA
<213> Streptomyces coelicolor
<400> 4

Claims

Method of providing a sensor surface with a compound comprising the steps of providing at least a part of the surface of the sensor with a hydrophobin coating and contacting the compounds with the hydrophobin coating to form a non-covalent bond between said hydrophobin coating and said compound.
Method according to claim 1 wherein said compound is smaller than 15000 dalton.
Method recording to claim 1 wherein said compound is smaller than 2000 dalton
Method according to claim 2 or 3 wherein said compound is incorporated in said coating.
Method according to any one of claims 1 to 3 wherein the compound is an electroactive compound.
Method according to any one of claims 1 to 3 wherein the compound is capable of being converted into an electroactive compound
Method according to claim 1 wherein said compound is larger than 15000 dalton.
Method according to claim 1 wherein said compounds comprises a proteinaceous substance.
Method according to claim 7 or 8 wherein said compound comprises an enzyme.
Method according to claim 7 or 8 wherein said compound comprises an antibody.
Method according to claim 7 or 8 wherein said compound comprises a receptor.
Method according to anyone of claims 1 to 11 wherein the compound is at least partly hydrophobic or comprises a hydrophobic anchor.
Method according to anyone of claims 1 to 12 wherein a second compound is bound covalently to the hydrophobin coating.
Method according to anyone of claims 1 to 12 wherein a second compound is bound non-covalently to the hydrophobin coating, through a third compound being an intermediate compound having affinity for the second compound.
A sensor provided with a surfaces at least a part of which is provided with a hydrophobin coating comprising a non-covalently bonded compound.
A sensor according to claim 15 wherein said compound is smaller than 15000 dalton.
A sensor according to claim 15 wherein said compounds is larger than 15000 dalton.
A sensor according to claims 15 wherein said compound is a proteinaceous substance.
A sensor according to anyone of claims 15 to 18 comprising an electrode.
A sensor according to claim 19 wherein the electrode is a glassy-carbon electrode, a glass electrode or a Thin Mercury Film Electrode.
A sensor according to anyone of claims 15 to 20 provided with an electro-active compound.
A method for determining a voltage difference in a solution comprising contacting said solution with a sensor according to anyone of claims 19 to 21 and recording a current.